Understanding emerging patterns of low fertility in middle-income countries is of essential importance. We demonstrate that the use of the P/F Brass methods in Brazil to adjust for a presumed underreporting of births has the potential to overestimate the country’s 2010 TFR by about 8%. Our preferred fertility-register-based estimate is 1.76, substantially lower the officially reported 1.90. This overstatement of fertility in official statistics has important consequences: compared to our analyses, for example, the UN World Population Prospects (UN WPP) overestimate recent TFR levels, and underestimates additional TFR declines during 2015-30, resulting in a projected 2050 population for Brazil that is 7 million larger and almost one year younger than projections using a 2010 TFR of 1.76. Several other Latin American countries are possibly subject to similarly upward-biased official TFRs that result from the use of the P/F Brass method in contexts of declining TFRs accompanied by an onset of fertility postponement. We hence believe several Latin American countries have progressed further in the transition towards low fertility than is reflected in official or UN WPP estimates. Our analyses also suggest that the further use of the P/F method in these countries should be carefully evaluated.